Go to JKU Homepage
A Mutual Exchange Extending Beyond Individual Disciplines
Artificial intelligence and medicine are two closely intertwined fields which thrive on discussions extending outside the confines of their respective disciplines, as well as on debate and mutual understanding between various areas of expertise. During the course of the evening, we hope to initiate and encourage these discussions. The focus will be on approaching issues, recent developments, and new challenges facing this interdisciplinary subject area and ultimately provide audience members with food for thought.
Erich Kobler (JKU Linz) opens the evening with his keynote address titled “Gadolinium Dose Reduction in Brain MRI and Model-based Deep Learning for Inverse Problems”. Philipp Vollmuth (University of Bonn) will deliver the second section of the keynote address.
Afterwards, there will be an opportunity to share ideas and chat as part of a casual, relaxed get-together.
Event: AI in Medicine
DATE
Wednesday, September 18, 2024
5:30 - 7:30 PM
A small reception and refreshments will follow.
LOCATION
YOU CAN EXPECT
A keynote address, discussions, networking opportunities
Keynote
Gadolinium dose reduction in brain MRI and model-based deep learning for inverse problems
Recent deep-learning-based approaches for reducing gadolinium-based contrast agents (GBCAs) in MRI face challenges in accurately predicting contrast enhancement and synthesizing realistic images. In this presentation, we introduce novel deep-learning methods, outperforming existing techniques across different scanners, field strengths, and contrast agents. In the second part, we combine variational methods and highly expressive deep-learning-based image priors by a mean-field optimal control problem. This combination allows for rigorous mathematical analysis including stability estimates and simultaneously leads to state-of-the-art results on various image restoration and reconstruction problems.
Erich Kobler
Erich Kobler, earned his BSc (2009-2013) and his MSc (2013-2015) degrees in Information and Computer Engineering (Telematik) and his PhD (2016-2020) in computer science at the Graz University of Technology.
After a PostDoc position at the Graz University of Technology, he accepted a senior lecturer position at the Institute of Computer Graphics at the Johannes Kepler University Linz. His current research interests include machine learning, computer vision, inverse problems, and medical imaging.
![[Translate to Englisch:] Erich Kobler](/fileadmin/_processed_/e/5/csm_pic_kobler_4acd7c7d22.jpg)
Clinical Translation of AI in Radiology - Opportunities and Challenges
Integrating AI in radiology opens up significant opportunities to improve diagnostic accuracy and streamline clinical workflow. The session will explore these opportunities, concentrating on advances in AI-driven image pre-processing, image analysis, and decision support, while also addressing the challenges to ensure model generalizability, stringent clinical validation, and real-world applications. Participants will learn more about state-of-the-art AI in radiology, focusing on neuroradiology as well as clinical applications and the critical steps required to successfully implement AI in a real-world setting.
Philipp Vollmuth
Philipp Vollmuth is a physician-scientist and full professor (endowed by the Else-Kröner Fresenius Foundation) for AI in Medical Imaging at the University of Bonn. He leads the division for Computational Radiology & Clinical AI at the University Hospital Bonn with a secondary appointment at the Division for Medical Image Computing at the German Cancer Research Center (DKFZ) Heidelberg. His research focuses on building and clinically translating state-of-the-art AI and big data analytics in the field of radiology. He is significantly involved in several national and international research projects related to AI in radiology. His scientific work has been published in leading journals (including Lancet Oncology, Lancet Digital Health, and Nature Communications) and has received numerous research awards.
![[Translate to Englisch:] Philipp Vollmuth](/fileadmin/_processed_/2/a/csm_500_hochformat_0f581ce3b7.png)
